Protocols for processing of tissue from arthroplasty infections vary and might affect the recovery of bacteria. We compared homogenization, bead beating and enzymatic disruption for recovery of live bacteria from tissue samples.

Suspensions of Staphylococcus aureus and Escherichia coli were prepared as controls. Three samples were taken from each and the first was bead beaten, the second homogenized, and Proteinase K was added for 10 and 30 minutes to the third sample before culturing. In addition, artificially inoculated pork tissue and known infected human tissue samples were processed by either homogenization or bead beating prior to cultures and results were compared.

Number of cycles of bead beating and homogenization and duration of Proteinase K treatment had significant effects. Bead beating for 2 and 4 cycles reduced the yield of S.aureus to 52% and 20% of control, and E.coli to 33% and 8%. Homogenization for 2 and 4 cycles reduced S.aureus to 86% and 65% of control, and E.coli to 90% and 87%. Proteinase K for 10 minutes and 30 minutes reduced the yield of S.aureus to 75% and 33% of control, and E.coli to 91% and 49% respectively. Inoculated Pork tissue showed a reduction in S.aureus recovery of 90% for bead beating compared to homogenization, and 80% in the case of E.coli. Bead beating of infected human tissue samples reduced the yield by 58% compared to homogenization.

Bead-beating is a common recommended method of processing tissue from arthroplasty cases. However, even though it produces a homogeneous sample, it does so at the cost of significant loss of viable bacteria. Homogenization and 10 minutes of Proteinase K incubation are almost equivalent, but the homogenizer is preferred being more controllable and cheaper. This should help to define guidelines for diagnosing infections using tissue samples.

Metal instrumentation (rods and screws) is used to stabilise the spine after trauma, malignancy or deformity. Approx 3% become infected often necessitating removal of metal. At surgery tissue samples and metal are removed for culture, but many clinical laboratories are not equipped to process metal or use simple culture methods. The causative bacteria exist as biofilms on the metal and they are often anaerobic and slow-growing, so conventional culture methods often fail to detect them. Also, they are common contaminants leading to diagnostic uncertainty. We have established a laboratory protocol to overcome these problems.

Removed metalwork was sonicated and the sonicate centrifuged and the supernatant discarded. Quantitative aerobic and anaerobic culture of the resuspended pellet for 14 days and microscopy were carried out.

Metalwork from 11 suspected infected cases was culture-positive (median 2857, 60–5000cfu/mL). Microscopy revealed an infection due to Candida albicans that would not have been detected otherwise. Bacteria were isolated from 8 of 10 non-infected cases (median 15, 0–35 cfu/mL). Conventionally processed samples failed to grow in 4 infected cases. (cfu/mL infected vs noninfected cases p=0.0093)

Micro-organisms on spinal metalwork grow as biofilms and they require sonication to dislodge them. The causative bacteria are slow-growing and P acnes is anaerobic and requires prolonged incubation. S epidermidis and P acnes are common contaminants and quantitative culture helps to distinguish pathogens from contaminants, removing the diagnostic uncertainty that conventional methods give. Microscopy of the sonicate can reveal micro-organisms that fail to grow on culture. We recommend that sonication of metalwork, prolonged anaerobic incubation and quantitative culture be adopted to improve diagnostic clarity for spinal instrumentation infections.